# "I want to come back as the bond market. You can intimidate everybody." --- James Carville (with the mandatory AI angle, of course) > Source: > Published: 2026-08-19 11:30:00+00:00 This is a big story, certainly bigger than you might have guessed from the coverage. Smart commentators like Patrick Boyle and the Financial Times have been ringing the alarm bell on the state of debt, particularly involving the unregulated Wild West of private credit, for a while now. It is one of those stories where the implications can genuinely keep you up at night. [Bond markets are getting hammered. Here’s what’s driving the sell-off] by John Towfighi Investors’ concerns over a range of issues from inflation to hefty government deficits are driving a bond market sell-off, creating a headache for policymakers and pushing up borrowing costs for governments and consumers. The 30-year US Treasury yield on Tuesday hit 5.34%, its highest level since 2007. The 10-year yield hit 4.74%, hovering near the highest level of President Donald Trump’s second term. The major factors here are inflation, tariffs, the accumulated bad habits of the era of free money and the asset bubble, national debts, the increasing loss of Fed independence, war, and the generally precarious state of the world economy. There's also, however, a secondary but definitely non-trivial AI aspect of all this. Government bonds are also under pressure from a wave of new debt from companies, including tech firms focused on artificial intelligence. Tech companies are issuing debt to fund the buildout of AI infrastructure, and those bonds are competing with government bonds for investors’ attention. Less demand for government bonds pushes prices lower, which pushes yields higher. “Hyperscaler borrowing to fund AI infrastructure is competing for the same pool of buyers at the same moment governments need those buyers most,” Nigel Green, CEO at deVere Group, said in a note. “Crowd two urgent borrowers into one market and the price of patience goes up for everybody.” The trillions of dollars that have been poured into the AI bubble are unprecedented. There is no comparable phenomenon in the history of the markets, at least in the modern age. To further complicate things, AI-related debt makes up a substantial number of the blocks in the Jenga Tower of private credit. To further exacerbate the problem, or at the very least add insult to injury, people like Fed Chair Kevin Warsh are actually using the promise of the as-yet-unobserved AI-driven explosion in productivity as an excuse for some possibly risky policy decisions. I'd argue the real tragedy here is that much, probably most, of these trillions of dollars have been wasted, not because large language models are without value, but because a combination of greed, false urgency, and pipe-dream delusions have driven a string of spectacularly bad decisions, including billions for worthless startups and trillions on what now appears to be a massive overbuild in data centers. A smaller, slower, and smarter approach would possibly have pushed back the rate of progress a little, but it would also have allowed for fewer gas turbines poisoning neighborhoods, better understanding and regulation cutting back on things like mass disinformation and AI psychosis, and a few less trillion dollars of debt threatening the world's economy.